> On Sep 21, 2023, at 13:06, Alexander Okonnikov 
> <[email protected]> wrote:
> 
> Hi Owen,
> 
> RFC 2328 doesn't require MTUs to be matched. It unambiguously says that 
> "Interface MTU ... is larger than the router can accept on the receiving 
> interface ...". RFC 2328 even doesn't say something about MTU of receiving 
> interface. So, if implementation on router A performs check for MTU mismatch 
> (i.e. from router A point of view they are expected to be equal), it is 
> misinterpretation of RFC 2328.

Perhaps… However, it’s such a widespread misinterpretation that if you do a 
search for “OSPF stuck in EXSTART”, almost every result is a discussion of how 
this happens when the interface MTUs don’t match:

https://www.google.com/search?client=safari&rls=en&q=ospf+stuck+in+exstart&ie=UTF-8&oe=UTF-8
🔎 ospf stuck in exstart - Google Search
google.com

> 
> Yes, RFC 5838 says something another - it, for unknown reason, says, that RFC 
> 2328 says that DBD to be rejected if Interface MTU value in DBD is greater 
> than the receiving interface's MTU. I don't know why it is needed to compare 
> those two values, but, again, even in this case there is no requirement for 
> MTU matching. There is requirement for to be not larger, and this is 
> different requirement. Please, correct me if I'm wrong.

Well… sort of… If A must not be larger than B and B must not be larger than A, 
then A must equal B.

True, the router with the larger MTU will accept the smaller MTU 
(theoretically, though FRR OSPFD does report the received smaller MTU in the 
DBD as a problem and won’t progress the negotiation and this is also true of 
many other implementations).

Hard to imagine how one could send a package complaint with a 0 MTU, so this 
doesn’t seem all that unreasonable absent special casing for MTU 0 beyond the 
“OSPF virtual-link” scenario intended in RFC5340 and RFC5838.

Owen


_______________________________________________
Lsr mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/lsr

Reply via email to